Is there any demand for a game asset focused modeling tool?
polycounter lvl 9
Hi everyone,
Recently things have been falling into place for me in terms of opportunities, connections and finance. I thought I would mention this issue as it has been around for a long time now. We have been shackled by programs that are jack of all trades and focused on other mediums (vfx, architectural, film etc). Our current options have been Max, Maya, Modo and Blender over the last decade. Other software like Silo and Nvil were sadly shortlived/unfocused.
It is generally a hotly debated topic on the choice of software for our industry and personal needs.
We are all aware of the bloated UI, unintuitive workflows, terribly bound shortcuts, difficult to learn, buggy, broken releases these software have. Adding 3 or 4 big features in a new yearly release that are completely unrelated to our needs is nothing short of depressing. A modeling ribbon, pedestrians, I mean comon?!
My proposition would be a new tool which would be completely game modeling/asset focused. Speeding up workflows and feeling fantastic to use everyday.
Some ideas come to mind:
I am interested to hear everyone's opinion on this topic
Recently things have been falling into place for me in terms of opportunities, connections and finance. I thought I would mention this issue as it has been around for a long time now. We have been shackled by programs that are jack of all trades and focused on other mediums (vfx, architectural, film etc). Our current options have been Max, Maya, Modo and Blender over the last decade. Other software like Silo and Nvil were sadly shortlived/unfocused.
It is generally a hotly debated topic on the choice of software for our industry and personal needs.
We are all aware of the bloated UI, unintuitive workflows, terribly bound shortcuts, difficult to learn, buggy, broken releases these software have. Adding 3 or 4 big features in a new yearly release that are completely unrelated to our needs is nothing short of depressing. A modeling ribbon, pedestrians, I mean comon?!
My proposition would be a new tool which would be completely game modeling/asset focused. Speeding up workflows and feeling fantastic to use everyday.
Some ideas come to mind:
- - Completely game asset creation focused.
- - Meant to supplement the already wide array of programs we use for sculpting/texturing/baking
- - Blazing fast modeling techniques, context sensitive, intuitive shortcuts and UI. Parametric modeling? Max's stack would be great here. UV mapping could be faster/more artist friendly.
- -Completely customizable UI and hotkeys, minimal to no bloat, drag and drop, shareable user configurations. Anyone remember perna's UI proposal for max? That was amazing.
- - As much automation as possible, one click operations (uving, test baking, exporting, previewing) or one click operations for our secondary programs/engines.
- - Open source maybe? Complete scripting language and user sharing.
- - Free, Indie, Professional versions.
- - Working alongside the game development community closely. Working with industry professionals on their needs and tools to speed up workflows. (recording actions and seeing where the most tedious actions are and eliminating them would be great)
- Immediate hotfixes, not becoming a bug ridden mess.
I am interested to hear everyone's opinion on this topic
Replies
Currently if you get something in Modo, Maya, Max... its like getting tossed a bone from the dinner table... I am all for seeing that change... assuming what is done actually can compete, in a game-centric manner, with three big M's (max maya modo).
what does 'game asset focused' mean exactly? the days of simple low poly are long gone and personally speaking i tend to use a fair amount of features originally intended for other tasks and repurpose them and bake the output down for realtime compatibility. would not want to be stuck in something with a single focus like silo again.
i would recommend having a look at blender and perhaps aiming to turn that into the app you envision? it's very interesting already but requires a good deal of customization to get it to anywhere we expect after using the commercial standards. plenty of opportunity to improve it's workflows, redesign the interface, add new tools and make it that strong, intuitive and automated package of your dreams. then release it as a stand-alone flavour/fork?
To me, the biggest issue with existing 3D apps is their half-baked customization options. Other then ZBrush none of them allow button-by-button drag/drop customization without scripting, only NVil allows context sensitive hotkeys (iirc), the apps that allow for custom navigation don't allow for stacked keys (except for 3D-Coat), custom sticky key support is spotty at best across the board, and you can completely forget about custom gesture-based controls.
Also about open source; I think it would be great to see Blender finally get some competition. There are tons of free 2D painting apps, but Blender is really the only decent one for 3D at all (unless you count Wings3D). If it was good enough, I would gladly contribute code to an open source 3D app, but I have no interest in doing so for Blender as I don't trust the Blender foundation (they seem to ignore or throw away a lot of viable user code).
Looks like 3D Coat is looking into adding modeling capabilities as well.
Yes, I believe the entire modeling(sculpting)/uving/texturing could be entirely done in a lightweight modeling package.
Heres a picture of my interface, mostly basic, but I added all the tools I need into a nice convenient bar, super handy.
https://hostr.co/file/QeVf3nLjuZSG/screen.png (that thing in the center to the right)
Also theres a nice folder structure with stacks, you can change and add new layouts with completely new interfaces and switch them. Its lacking some of the raw modeling power 3DS has, but to expand on its perfect. Also very stable and fast.
You can also change fonts, UI colors for really everything, every function that exists can found in a nice searchable list, can be rebinded and made text buttons or icons off. Sadly the baking tools are super basic, else you could go from subdivision modeling straight to sculpting UV and baking straight in C4D without ever leaving the software.
Generally it would be probably best to expand on an existing package that is no patchwerk monstrosity, making tools for Modo or so would probably be the best bet to get the package you want.
On the contrary, Blender is super stable and free - one guy from our team use it at work for more than 2 years with no problems at all. This is life`s black humour - open source software is more stable than paid ones..
If you have to ask this question then the answer is no. Why? Because if you're trying to revolutionize a particular part of the game development pipeline... You need to already have a strong vision that you know for a fact will do just that.
I do believe there is a market for this kind of thing. But if you're going to be the one to do it then you need to make sure that what you're building will be the greatest thing since sliced bread.
Substance painter is a great example in the texturing field of what you are talking about doing. It takes everything a game artist needs to do and makes it fast and fun to do.
I'll include a video because its actually quite relevant to this thread's topic, as I think they have kind of hit the nail on the head in regards to a game-centric work space and work flow (still needs more polish but huge leap forward).
Now imagine having that in conjunction with the Zen UI: http://www.mechanicalcolor.com/modo-kits/zen (which shows what kind of enhancements the users have created for themselves). The worst offender in Modo really is the right side of the screen (attribute/property panels) and lack of color customization, but thats really about it... and a bit if it has been helped by that games layout.
That said Blender's UI is actually a mess, not that I dont like the way it "feels" but its not organized very well at all. Its also over complicated and un-intuitive in regards to being unconventional. One of the worst things they did was make the mouse location matter when trying to work with the UI. If its hovering over one editor, its own keymap takes over. This means you have to juggle many keymaps depending on where you park your mouse. Kind of odd really. It too can use some stream lining. If it was that great I would still be using it btw.
Regarding stability.... I agree. Modo took quite a hit with 901. The reason for this was their approach to development. They tried to cram too many changes in all at once before a launch that there was just no way the QA could catch all the problems and have them be dealt with. So with Modo 10 they evolved/adapted. Every version will have a set of new features and updates rolled out over the course of a year. Right now its set to 3. This means they dedicated their development to one set of features, generally in a particular area and build upon that. It lets them get more done in a more drastic way while also getting rid of the problem that resulted in instability at launches and the subsequent catch up (removing the issues) that follows. I cant say too much but the next update for 10 is something a lot of Max and Blender users appreciate in their own application.
I dont quite believe your "one guy" had no problems... but he probably worked around issues that wouldnt be a problem in other applications.. such as FBX exports. Stability wise though, Blender is a tough nut to crash... i have seen maybe 1-2 crashes with it in my lifetime and that was due to using their development builds.
Anyways, Modo has free demos for 10 now so no more excuses not to test it out (always recommend turning off trackball rotation and setting a familiar navigation style such as Maya within its settings).
Center toggles change the viewport layout or whats within that space, so center makes sense. Toggles on the right hide or show the object/shader tree..ect, plus pulls up panels. So its all segmented properly. The left panels can be closed too just by dragging them into pullout buttons or a floating properties panel. I cant think of a more logical way to lay it out, can you? Keep in mind also this is within their existing UI frame work, so the turn around for developing it was probably a lot shorter than if they started rewriting things that might end up causing far problems. If this kind of streamlined layout is popular, it can easily drive future dev with that in mind.
Centered tabs, for or against, shouldnt be a fashion statement for UIs btw. It works or it doesnt. Is it accessible, does it convey what area its correlates to? For example, with the current layout I can just do this to maximize the work space. You can see some custom tools on the upper right, which can really be dragged around and docked anywhere. Also have a custom quick boolean and snap bar I could slap in there. Its really not that bad.
Personal layout: https://drive.google.com/file/d/0B6C-cBuP60GjOUFfOEh0M0M2Z0k/view?usp=sharing
The function for me right now is more important than the aesthetic, which in that case I prefer things to be more dark, boxy and smaller in scale...
I think "replacing" any package is much harder than creating a great package which is good at collaboration with other packages. Who doesn't hate packages which are in your workflow but just don't get along because the developers thing theirs is the only one true solution?
Personally, I'd rather have better interconnection between those programs than yet one more app to add to the mix. Each new tool in the pipeline adds to the mess of wasted time (and introducing another point for errors) due to data exchange (but also training, maintenance, support). Imagine you could send data from any app - Painter, 3D Coat, Max, Mudbox, ZBrush - as easily as pressing a button, without losing everything, without having to flatten your textures and without losing modifiers and history. That would save so much time. I've seen some clever studio proprietary solutions during the years and those were some of the biggest time savers, because they allowed artists to move between specialized apps easily and consistently, so they could use the strongest tools at their disposal at any time.
In any case, stay focused. Specialize. Be open and make sure the tool works well with anything else that's in the pipeline.
1. Auto UV
2. Auto Re-topo
3. Auto Bake
Let me use any other package for modeling or texturing.
Dataday said:
There are problems with Blender regarding export, as with any other app out there (AD and modo included). Export from app to app is the main problem with all 3d pipeline: orientation gets rotated, smoothing gets killed, UVs welded, rigs get broken, animation baked, materials removed - the list goes on and on.
@PolyHertz
I completely agree with you there. I have yet to see an app that can fully let you drag and drop and build out your own custom UI like modular lego pieces. That is one idea our group talked about. One example of annoyance is the 3dsmax command panel, it is quite common people will drag it out into 2 panels, while 80% of the tools there are unused and bloating screenspace. The program being open source opens up so many possibilities with users implementing/submitting features on their own. We could then implement those into the code so each new release does not break that persons implementation. I know this is extremely troubling when finding custom tools for blender which become broken shortly after an update with no guarantee of support.
@Kwramm
I completely agree with you here. In fact the cross program automation is something we've been thinking hard about ever since we heard Mike Pavlovich's automation talk at GDC. It can be done but it's just a rarity anyone has publicly release something to save everyone's time. One click operations and the closer we are to a "make art button" is what we should be moving towards. We need to let artists be artists and computers do the tasks we shouldn't be doing. It needs to stop being limited to technical knowledge and instead limiting the artwork to the artist's creativity. For example hard surface art imagine you could entirely make your high poly model out of kitbashed pieces for boolean/mesh fusion operations at a few key presses. High poly done. Then a one click operation to create a low poly cage from this high poly with perfect topology, and all that has to be adjusted is a slider for you're true vert count/tri limit.
Then a one click operation for generating UVs where it would find the best angles/smoothing groups, you could define if the asset is being viewed at a certain angle. With algorithms it tries to find the best possible solution of eliminating all seams from that view/perspective. Or combining as many islands as possible in consideration to potential bake errors/smoothing. Heck generating low poly cages, UVs and test baking calculations based on the aforementioned could be done in a single operation. It's already been halfway done by that recent custom blender tool. It's already close in this area with regular flatten mapping and some basic scripting. Now have the programmers create the algorithms to support this basic function and we'd get closer. What mattered in the entire process is how you created the hard surface work, does it match the concept/pics in proportions/scale? Or how was the design handled, were there any changes or improvements? This rings true to traditional 2d artwork.
If this were the case why aren't people up in arms of substance tools not branching out into other mediums? I would consider Allegorithmic's products to be a great success in this industry, they lead by example considering being used in most AAA studios these days. I would not call Modo's UI/tools to be anywhere near the flexibility and access that I am describing. The video you've linked hasn't exactly impressed me at all and seems quite rudimentary in terms of industry support. With any jack of all trades tools there simply won't be focus where it counts.
Even mesh fusion cannot be used in production and wasn't at all what it was hyped up to be. I think something like mesh fusion which delivers perfect topology could have been a game changer in hard surface work. It would save an incredible amount of time and being a focused tool people could be allocated day in and day out to make sure that is in a future update in a matter of months (not years like quad chamfer coming to 3dsmax).
@thomasp
With a game asset focus this would be the driving factor in what features and elements would be included with each update. There would be no waste of tools here and everything would serve a fundamental purpose for a game artist. All forms of environment, characters, hard surface tools would be at the user's disposal in each and every update. Fluid high poly modeling, flexible retopo/low poly cage abilities, modular kit building and extreme automation of mundane tasks are important to this industry.
Current apps are incredibly industrial and a focused tool for these options hasn't quite existed yet. For example substance saves time by allowing you to paint in multiple channels. Previewing results is instantaneous rather than setting up photoshop actions and alt tabbing to see results in-engine. The days where people would spend an entire day texturing a weapon are long gone and reduced to a mere couple of hours. There is no denying what substance has done as being dedicated for one purpose has really pushed this industry forward and saved thousands of hours.
I think its that people get uncomfortable when the idea of possibly moving away from their already time and money invested in applications.
Think about it. We pay for updates, or version of software that focuses on features we will never use. All we want are those game related workflows and the innovation based around that. Yet we go through this waiting game... waiting for a damn bone to get thrown our way. A lot of the people working on the very software come from those film or vfx backgrounds, which is why these game features feel so ...awkward. There is no real "understanding", but we accept what we are given and follow it up with a "please sir can I have some more".
Its pretty messed up. Wanna know why Allegorithmic/Substance did well? They said "f**k it, we are approaching this from a games point of view and looking at both the problem and the solution". Bam! Before we might have said "eh we dont need it... we dont want it" but now everyone and their second cousin is working with Substances.
My point is, these people some times dont know they want something or need something until its there. If you need to pave the road, show the game artist that they are treated like a 3rd wheel, then do so. Dont feel disheartened. Innovate, focus, say "you know what, game dev could use this and its so much freakin better" with no regrets. That kind of thinking starts to become contagious. Thats when Autodesk, The Foundry, and others start trying to catch up to you. Pixologic is a great example.
There is innovation to be done, a focus or a home to be built for this workflow. Some might not just know it yet. The market is there for the taking, its just up to you guys to see if you have what it takes to match whats expected and then surpass it.
Hmmm, I'd say the general consensus is, Bring It On ! Everyone is willing to adopt the next big thing. It just has to be good.
It is generally a hotly debated topic on the choice of software for our industry and personal needs.
It only tends to be hot topic when some sort of brand loyalty is involved. When looking at things objectively the flaws and advantages of various software packages are quite straightforward to evaluate.
Look at how quickly the substance tools have taken over from the status quo of photoshop.
I would love to drop the overpriced bullshit that is Autodesk, in favour of a more directly tailored tool that actually wants to innovate and simplify the way we work.
@tynew
Sorry if my previous post seemed a bit negative. I'm definitely open to new packages and would be extremely excited to try out something new. I think asking the internet about this kind of topic can be a bit dangerous. Sometimes people don't know what they want until they see it. If you have an awesome vision I would say... don't let anything stop you. I would love to be free of autodesk, so by all means... go and make something epic.
If there was a stand alone package that did this kind of thing and was properly marketed I believe it would be quite successful.
http://polycount.com/discussion/164254/maya-surface-based-asset-generation
I'm a Modo user and I'd buy a new app that purely focus on game art! There is still a market for this IMO. Bring it! Do you guys have a roadmap?
It will be extraordinarily difficult to do what you're describing, avoiding the "bug ridden mess" and other features (that admittedly most modelling apps do just fine), but wish you luck!! If you have great ideas, don't ask for permission, do it and get people to need your ideas
which is exactly why i suggested looking at it earlier as well. the UX could be reworked, the toolset streamlined, new bits added. functionality wise it is already a good package that is very extensible (switched to from 3ds not too long ago myself) but remodeled with a good vision it could actually become the best choice. it also has an audience with plenty of people who are already well versed in writing addons and likely with a willingness to join a new community.
add commercial support for studios down the line to make it 'legit' and you might really have something there.
as opposed to doing it all yourself and going through the entire software dev baby step cycle which probably means nobody sane will consider it for real before version 3 or so.